1. Field of the Invention
The present invention relates to systems and methods for the spectral efficient transmission of signals, and more particularly, for combining transponder bandwidths for source and forward error correction encoding efficiency.
2. Description of the Related Art
Digital Direct Broadcast Systems (DBS), such as DIRECTV®, which is provided by the assignee of the present invention, have become very successful. However, as such systems evolve, there is an increasing demand for additional bandwidth and/or more efficient use of existing bandwidth to carry an ever-increasing set of audio, video and data services.
Transmission of video signals over fixed capacity transmission channels poses a number of technical challenges. Raw, uncompressed video requires bandwidth that can exceed the available bandwidth on a satellite channel by orders of magnitude. Video is therefore compressed prior to transmission, using techniques based on standards such as MPEG-2 and MPEG-4|H.264. These techniques can greatly reduce the required bandwidth while maintaining high visual quality. They do result, however, in a compressed bit stream that has a variable bit-rate. This is because the amount of compression achieved greatly depends on the actual content of the video: scenes with complicated texture or fast moving objects are compressed less and vice versa.
Typical wireless transmission channels, on the other hand, have a fixed transmission rate. To accommodate a variable bit-rate compressed video signal, the fixed wireless channel rate must exceed the peak bit-rate of the video signal. If the peak bit-rate of a compressed video signal exceeds its average bit-rate by a significant amount (usually the case with most types of content) there is a significant waste of the wireless channel capacity. This is typically a precious resource, especially when the wireless transmissions are made via satellite. For example, if the average video bit-rate is 10 mbps and its peak bit-rate is 25 mbps, it would require a satellite channel with a capacity of at least 25 mbps, but on average, 15 mbps of capacity will be wasted.
To reduce this waste of transmission channel capacity, multiple compressed video signals are usually multiplexed together using a statistical multiplexer (statmux) prior to transmission over the fixed bit-rate satellite channel. Since the peaks in the bit rate of one media program signal are unlikely to occur at the same time as the peaks in the bit rate of a signal for another media program, the proportion of the channel capacity needed to account for bit rate variance is reduced. This allows the bit rate of the combined signal to be more closely matched to the channel capacity, thus reducing waste of transmission capacity.
According to the central limit theorem , the variance of the sum of N random variables is the sum of the variance of each random variable. Hence, as N increases to include more random variables, the variability of the sum of those random variables decreases. Considering the instantaneous bit rate of a video stream to be a random variable with a variance and a mean, one would expect the variance of the instantaneous bit rate of N video streams to be a smaller proportion of the average bit rate as the number of video streams goes up.
For example, suppose a signal comprises the sum of five signals, each having an instantaneous bit rate with an average of 10 mbps and a standard deviation of 5 mbps (hence, a variance of 25 mbps). Since there are five such signals, the average bit rate for the combined five signals will be 50 mbps, and the variance of the combined five signals will be 5*25=125 mbps, which equates a standard deviation of about 11.2 mbps. As can be seen from this example, although the average instantaneous bit rate quintupled, the standard deviation of that instantaneous bit rate increased by only a little more than a factor of two. Since the channel capacity must accommodate the peak bit rate, combining five such signals reduces the overhead required because the variance of the combined signal is reduced. This is because while one signal may have an instantaneous bit rate well above its average, another signal may have an instantaneous bit rate well below its average, thus allowing transmission of both signals with little or no overhead.
For an even simpler example, consider a signal with five video signals having the above characteristics (10 mbps average bit-rate and a peak bit rate of 25 mbps). If those signals are multiplexed such that the 25 mbps peak bit rate in one signal can be combined with 5 mbps bit rate troughs in two signals and 10 mbps average rate of the other two signals, the total bit rate capacity requirement is 55 mbps (5*10 mbps average rate plus [15−(2*5)] to account for the peaks), which is only 15 mbps (or 30%) more than the sum of the five average bitrates (50 mbps).
In practice, additional bandwidth is required for audio, error correction and transport signaling overhead, but the foregoing simple examples illustrate how statistical multiplexing can reduce the waste associated with transmitting a variable bit-rate compressed video signal over a fixed bit-rate channel such as that of a satellite transponder. This reduced waste can be used to increase channel capacity.
If the capacity of the fixed bit-rate channel is not significantly greater than the average bit-rate of the compressed signals (say at least four or five times), then little will be gained by statistical multiplexing. With the advent of high-bit rate content such as 4KTV, full resolution 3DTV and even 1080p60 content, existing satellite transponders will not have the capacity of transmitting several such streams, even if they are statistically multiplexed together. There is therefore the need to find a solution to maintain the gains of statistically multiplexing several such high bit-rate compressed video signals while at the same time using the existing satellite transmission capabilities.